<quote who="Ross Golder"> > Just dug this suggestion from Stric up. Don't know if it makes any sense > in the context of the stuff you're doing...
(Snip good suggestions about reducing the number of mails hitting Amavis.) So, I've *massively* reduced the number of mails going to Amavis simply by reducing the number of mails being allowed to go through the MTA. It looks like roughly 1 mail per second going through Amavis, which correlates with the number of received mails per minute (mpm) on mailgraph [1]. We can handle 75mpm, *internally*, no sweat. Realistically we can handle a heck of a lot more, even before we give Amavis more head room (cranking up its max_servers/max_requests, etc). So, certainly something to think about in the future, but now that we have good information about the behaviour of our mail load, we need to watch it for a while, to get a better grip on it over a useful period of time. Then we can start thinking about further optimisations. (Replying on gnome-infrastructure for temporal documentation purposes.) Thanks, - Jeff [1] http://mail.gnome.org/mailgraph.cgi - now with bonus virus/spam stats. -- linux.conf.au 2007: Sydney, Australia http://lca2007.linux.org.au/ "Linux continues to have almost as much soul as James Brown." - Forrest Cook, LWN _______________________________________________ Gnome-infrastructure mailing list [email protected] http://mail.gnome.org/mailman/listinfo/gnome-infrastructure
